Improving online video

I’m not such a big fan of predictions like we see them all over on the web at the start of the New Year. There are a few good ones like those of John Batelle for instance, but other than that many predictions sound more like acknowledgement of something that’s already happening today or a wish-list for something we would like to see happen.

One good example is mobile for instance. It’s going to be the year of mobile since 2005 I think and as I said before, 2008 is not going to be the year of mobile either. Or talk about video, I had a discussion with someone recently about online video after a statement that 2008 would be the year online video will get big. That’s just not true, 2007 was that year already. As a counter argument I did say I believe 2008 will be the year in which online video will become more useful, of better quality, with better metrics, better advertising, … you name it.

And that’s what is happening today indeed. Yesterday Read/Write Web reported on the launch of Dailymotion HD upload & playback, including automatic bandwidth detection which allows easy switching to lower quality versions. Check out the HD example in their post. Now I can see how HD quality might not be on top of everybody’s wish-list for online video, but I do believe it’s a must have for future development of online video.

About a week ago, MIT AdverLab reported on a new innovative technology related to video advertising. The technology developed in Microsoft’s AdCenter Labs included tools for content analysis and speech recognition for advanced contextual advertising. (again disclaimer: Microsoft is my employer). Definitely take a look at the video below to get an idea of what they exactly mean with that:

Now as I am in advertising, I’m interested in this but not only for advertising purposes. If you watched the video you will understand that there are also opportunities for websites for instance to relate archived content to in-video content (think news sites) or how the technology can create automatic chapters in a video for more useful video browsing for instance. There are quite a few exciting technology usages you can think of with this new development.

A last improvement will be about making video more searchable. The technology mentioned before will definitely be able to help in that area, but quite a few startups are working in that area as well. A few weeks ago at the LIFT Venture Night we saw Viewdle showing off their approach on this, including face recogniation etc. Neat stuff as well.

Content analysis, speech recognition, chaptering, contextual advertising and content (based on the video – not on title or tags), HD, … Yes, online video is big already, 2008 will be the year in which it will get better.